Abstract
ABSTRACT
Background
Burn severity assessment traditionally relies on clinical and injury‐related factors. However, the influence of social determinants of health on burn outcomes remains poorly understood. We aimed to develop a multidimensional prediction model of burn severity, defined as intensive care unit (ICU) admission, incorporating both clinical and social factors.
Methods
A retrospective cohort study of burn patients aged ≥ 15 years admitted to a specialized burn center in Cali‐Colombia between January 2023 and December 2024. Sociodemographic, clinical, and injury‐related variables were collected from institutional records. The primary outcome was ICU admission. Variables associated with ICU admission were evaluated using multivariable logistic regression, and model performance was assessed through discrimination and calibration metrics.
Results
A total of 160 patients were included, of whom 68 (42.5%) required ICU admission. Median age was 40 years (IQR 30.5–51.5), and 77.5% were male. In the multivariable model, greater total body surface area burned (OR 1.11, 95% CI 1.06–1.15;
p
p
= 0.030), and subsidized health insurance status (OR 3.50, 95% CI 1.38–8.87;
p
= 0.008) were independently associated with ICU admission. In the Colombian healthcare system, subsidized insurance primarily covers low‐income and socioeconomically vulnerable populations. The model demonstrated excellent discrimination (AUC 0.892) and good calibration (Brier score 0.128).
Conclusions
ICU admission following burn injury is influenced by a combination of injury severity, physiological response, and social vulnerability. Incorporating social determinants of health into burn severity assessment may improve early risk stratification and provide a more comprehensive framework for predicting critical care needs, particularly in resource‐limited settings.